DB
Get started
Back homeIndustry

Telecom

Flow records, windowed cell KPIs and the live state of every element, in one engine for network telemetry.

The problem

Where the usual stack breaks

  1. 01

    Network telemetry is relentless: flow records, interface counters and radio KPIs arrive continuously from thousands of elements, and the interesting questions are about right now. Which cells are degraded, which links are saturating, which elements went quiet.

  2. 02

    The common stack splits this across a stream processor for windowed KPIs, a cache for element state and a warehouse for history: three systems, three schemas, and reconciliation glue that itself becomes an outage source.

  3. 03

    You want one engine that keeps the history durable, maintains per-element current state, and computes windowed KPIs as the records arrive.

Where NYXDB fits

Use-case journeys

01

Flow & counter ingest

Durable columnar history of flow records with keyed element state.

IoT & telemetry
02

NOC observability

Raw telemetry becomes live per-cell metrics, maintained by materialized views.

Observability & logs
03

Real-time dashboards

Dashboards read maintained relations and subscribe to live changes.

Real-time dashboards

Architecture

How NYXDB fits network telemetry

Flow records and KPIs land in one table; keyed tables hold the latest state per element; STREAM SELECT over HOP or TUMBLE windows emits each KPI aggregate to NOC dashboards when its window closes.

  1. 01

    Ingest

    Flow and counter records arrive through SQL, write batches or native bulk frames.

  2. 02

    Element state

    The latest row per (element, metric), in an UPSERT table.

  3. 03

    Windowed KPIs

    KPIs aggregate by event time over HOP windows in STREAM SELECT, released once with EMIT AFTER WINDOW CLOSE.

  4. 04

    Live streams

    Element-state changes reach NOC dashboards as they commit, with STREAM SELECT.

In practice

Representative query

The examples use the same two tables on every page. See the SQL reference for full syntax.

Transfers per minute, as a windowed query
STREAM SELECT window_start, window_end, token, COUNT() AS transfers
FROM TUMBLE(transfers, transferred_at, INTERVAL 1 MINUTE)
GROUP BY window_start, window_end, token
EMIT AFTER WINDOW CLOSE;

Capabilities

What you get out of the box

Bulk ingest routes

Write batches and native bulk frames.

Keyed element state

Latest row per element.

Windowed KPIs

Tumbling, hopping and session windows in STREAM SELECT.

Live push

Changes reach NOC dashboards as they commit, with STREAM SELECT.

Learn more

Go deeper

Streaming

Windows and EMIT.

Tables

Current state per element.

Streaming reads

Snapshot then incremental changes.

Telemetry your NOC can query

Explore the telemetry solution or run a node.

Get startedRead the docs